Sam Altman is spreading OpenAI's money far and wide to keep the AI race alive (www.businessinsider.com)

OpenAI announced a multi-year hardware deal with AMD to deploy chips that could require up to six gigawatts of power, plus an option to acquire tens of billions of dollars in AMD stock. The agreement joins a recent flurry of infrastructure pacts — Nvidia committed access to 10 gigawatts of high‑powered GPUs alongside a reported $100 billion investment in OpenAI, OpenAI plans to co‑develop its first AI chip with Broadcom next year, and Oracle agreed to provide AI compute in a deal reportedly worth more than $300 billion over five years. Collectively these contracts lock in massive capacity for model training and inference and cement OpenAI’s role as a major demand driver for the AI chip market. Technically and strategically, the move signals a deliberate diversification away from Nvidia’s GPU hegemony to build supply resilience, spur competition, and potentially drive down costs through alternative architectures. For AMD, the deal validates its AI roadmap and could translate to “tens of billions” in revenue; for the broader ecosystem, it increases multi‑vendor support for large models and mitigates single‑supplier geopolitics and supply‑chain risk. The catch: these bets hinge on OpenAI’s continued growth — the deals are powerful demand signals that could accelerate infrastructure buildout, but they also concentrate market risk on OpenAI’s long‑term success.